A novel computational approach extracts more insights from maps that illustrate the clustering and distribution of galaxies across the universe compared to earlier methods.
A study spearheaded by the University of Michigan aims to advance cosmology by maximizing the effectiveness of telescopes and other instruments used to explore some of the universe’s most profound questions.
The initiative demonstrated that this innovative computational technique can derive more data from widespread maps depicting how galaxies cluster and weave through space.
Researchers are currently utilizing instruments such as DESI, the Dark Energy Spectroscopic Instrument, to produce these maps and explore the enigmatic properties of dark energy, dark matter, and various cosmic phenomena.
The obscure aspects of cosmology
Even as DESI captures public attention, it is recognized that further advanced tools will be necessary to unearth the answers sought. While some scientists develop the next generation of instruments like DESI, Minh Nguyen and his team are concentrating on refining our comprehension of the data we possess both now and in the future.
“As we transition to bigger and better telescopes, we might also inadvertently discard valuable information,” stated Nguyen, who played a key role as a Leinweber Research Fellow in the U-M Department of Physics. “We can enhance our insights while expanding our data collection.”
In collaboration with colleagues from the Max Planck Institute for Astrophysics (MPA), Nguyen employed a computational framework known as LEFTfield to improve how scientists interpret the universe’s large-scale structure.
“In the universe’s infancy, the structure appeared Gaussian — similar to the visual noise from old television screens,” Nguyen explained. “However, due to the interactions between dark energy and dark matter, the universe’s current large-scale structure is no longer Gaussian. It resembles a spider web.”
Dark energy is the force driving the universe’s expansion, yet it cannot be directly observed, which is why it is termed “dark.” The matter in the universe counteracts this expansion through the attractive force of gravity.
This matter exists in two distinct forms: the observable regular matter and the elusive dark matter, again highlighting the “dark” aspect.
A further compelling detail is that the vast majority of the universe’s mass and energy is tied up in these mysterious dark components. By analyzing cosmic maps, scientists can gain new insights into dark matter and dark energy, which play significant roles in shaping the universe’s web-like structure.
Using LEFTfield, Nguyen and his team demonstrated their ability to extract additional information from existing cosmic maps. Their findings were published in the journal Physical Review Letters and earned them the prestigious 2024 Buchalter Cosmology Prize.
To obtain this supplemental data, the team did not simply enhance the existing standard methods, which have proven invaluable. Instead, they adopted a fundamentally different strategy.
A fresh perspective with LEFTfield
The primary distinction lies in how LEFTfield processes data compared to conventional methods.
“Standard analysis methods require compression of the data,” Nguyen said. “This simplification aids in analysis and theoretical predictions, but results in some loss of information.”
In typical analyses, researchers employ computational models that categorize galaxies into pairs or triplets, facilitating more efficient statistical evaluations and calculations.
Nguyen mentioned that these methods are effective for the universe’s more Gaussian features. However, he and his colleagues recognized an opportunity to deepen the understanding of our non-Gaussian universe by keeping all the information that standard methods may overlook due to compression.
The innovative procedure, also referred to as field-level inference, treats cosmic maps as three-dimensional grids. Each unit cube, or voxel, serves as an individual data element that retains uncompressed details about galaxy density and distribution.
This approach maintains the integrity of the data in ways that conventional methods cannot achieve, Nguyen explained.
“I am enthusiastic about field-level inference because it fundamentally aligns with our primary goal,” stated Shaun Hotchkiss, host of the online seminar series, Cosmology Talks, which recently featured Nguyen and co-author Beatriz Tucci, a doctoral student at MPA.
“If we can measure the density field, why reduce the information within it?” Hotchkiss questioned. “Field-level inference is admittedly more challenging, but that shouldn’t deter Bea and Minh—nor the broader community.”
To evaluate LEFTfield’s effectiveness, the team calculated a cosmological parameter known as sigma-8, which essentially measures the universe’s clumpiness, Nguyen explained.
Compared to standard methods, the LEFTfield approach improved the determination of sigma-8 by a factor of 3.5 to 5.2.
“That’s akin to advancing from DESI to its next-generation successor,” Nguyen remarked. “Typically, progressing between two generations of surveys can take 10 to 20 years.”
However, before making that significant advancement, there remain challenges to overcome. A crucial step will be to integrate LEFTfield with specific instruments while ensuring it effectively accounts for noise and quirks in the data as it is collected, Nguyen stated.
Nonetheless, he remains optimistic that this approach will serve as a formidable tool.
“It accelerates our ability to glean insights into dark energy, dark matter, and general relativity—the foundational theories behind it all,” Nguyen concluded.
The research team included Fabian Schmidt, a cosmologist and group leader at MPA, as well as staff scientist Martin Reinecke and Andrija Kostić, who contributed as a Ph.D. student and later as a postdoctoral researcher.
Nguyen has recently completed his fellowship at U-M and is now a research fellow at the Kavli Institute for the Physics and Mathematics of the Universe in Tokyo.
